JPH0347704B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a new and improved apparatus and method for evaluating the condition of a specimen surface. The invention is particularly suitable for applications in the semiconductor industry. For example, the present invention can detect very small amounts of debris on a specimen surface. In addition, the present invention provides ion implantation adjacent the specimen surface prior to annealing.
The concentration of doped substances can be measured.

There is a great deal of interest in developing techniques for accurately analyzing specimen surface conditions. This interest is particularly important in the semiconductor industry, where there is a need to realize equipment that is not only extremely accurate, but also capable of evaluating extremely thin surfaces.
Especially strong.

One particular area where this need remains unmet is related to residue detection. More specifically, the etching and lithography processes used in the manufacture of integrated circuits are nearly undetectable, yet can lead to premature circuit failure. A thin layer of residue is left behind.

Referring to FIGS. 1 and 2, typical steps associated with semiconductor manufacturing are illustrated. More specifically, as shown in FIG. 1, an oxide layer 12 is formed on a silicon substrate 14. The next step in semiconductor preparation is to etch a pattern into the oxide by lithography and etching processes. For example, a layer of photoresist material 16 is applied to the upper surface of the oxide. A mask is placed over the photoresist material and light is applied. The unexposed portions are removed, forming the structure shown in FIG. Thicknesses are exaggerated for clarity.

Ideally, the above structure is then "etched"
However, this step removes the oxide layer in the areas where the photoresist was removed. A typical structure formed after this etching step is shown in FIG.

Unfortunately, mass production techniques do not always produce complete devices. For example, the steps to remove portions of photoresist are often incomplete.
Further, residue may remain at the position shown at 16A in FIG. Where residue remains, no step is taken to remove portions of the oxide layer. Etching the oxide layer where the photoresist has been properly removed can also result in lost sales. As shown in FIG. 2, residues of the oxide 12A may remain. In either case, integrated circuits manufactured from these imperfect structures are either inoperable or prematurely fail.

So far, industry has not yet found a good way to detect these residues. In most manufacturing facilities, integrated circuits are inspected by humans using microscopes. This is clearly very time consuming and inefficient. Moreover, the thickness of the residue, which has an undesirable effect on the manufacture of integrated circuits, is so thin that it is barely visible even through a microscope.

Another method is based on the principle of using fluorescence. In this method, the photoresist is illuminated with ultraviolet light,
The resulting fluorescence is observed. In some cases, residues of photoresist material can be detected. However, like standard visual inspection, fluorescence provides only a rough indication of whether the removal or etching step of photoresist material was successful or lost. In addition, fluorescence methods can only be applied to photoresist residues and not to other residues.

Another problem with these prior art detection methods is that
These are not very accurate or precise. More specifically, residual debris is often trapped at the bottom of the etched "hole" or "contact" as shown in the cross-sectional view of FIG. To detect these residues,
A system that can converge accurately is required.

During the manufacture of integrated circuits, it is often desirable to evaluate the surface condition of a specimen. for example,
Implantation of dopants with different electrical conductivities often imparts semiconducting properties to silicon substrates. Referring to FIG. 3, a silicon substrate 18 is shown having doped ions 19 implanted therein adjacent the upper surface of the substrate.

A typical implantation process involves rastering the ion beam onto the specimen surface. Some of the ions in the ion beam implanted into the specimen are introduced into the silicon lattice structure. The doped ion density in the specimen is related to the time the beam is focused to any point on the specimen surface. The depth through which ions penetrate is related to the voltage of the ion beam. However, even at maximum voltage, the depth through which the ions penetrate is relatively small, and the ion density is substantially at the upper surface of the specimen.

Ions implanted into the silicon surface are at interstitial locations, physically disrupting the lattice structure of the material.
In this state, the material does not exhibit desirable semiconductor properties. To overcome this problem, "activating the dope" is necessary in subsequent manufacturing steps. The dope material is activated by an annealing process. In this annealing step, the material is heated such that the lattice is repaired and ions move from interstitial positions to substituted lattice positions. In this process, doped ions replace silicon at various points in the lattice.
This annealing stage serves to remove defects in the lattice and release the electrons of the doped ions to conduction current.

Once the annealing step has taken place, the implanted doping level can usually be measured by known electrical resistance methods.
Since doped silicon is at least partially conductive, the doping characteristics are easily obtained by measuring electrical parameters. In the normal case where boron ions are implanted into silicon,
Current electrical testing methods can detect concentrations as low as 10 ¹³ ions/cm ² . Unfortunately, it is desirable to detect concentrations as low as 10 ¹¹ ions/cm ² .

Additionally, and more importantly, prior art electrical methods have another drawback of not being able to detect ion concentration prior to annealing. As mentioned above, prior to annealing, the electrons of the doped ions in the specimen are fixed at interstitial positions within the lattice structure. Electrical measurement methods therefore provide no information. In fact, there is currently no method to directly evaluate the dope ion concentration in a specimen before annealing. Evaluation at such an early stage is highly desirable as it provides information on the success or failure of the implant before incurring the expense of annealing. Furthermore, such information is valuable in the design of manufacturing stages for improving and modifying the characteristics of integrated circuits.

It is therefore an object of the present invention to provide a new and improved apparatus and method for evaluating the surface condition of a specimen.

Another object of the invention is to provide a new and improved apparatus for evaluating surface conditions based on the principle of reflectance.

Yet another object of the invention is to provide a method and apparatus for characterizing very thin surface areas, less than 100 angstroms.

Yet another object of the invention is to provide a new and improved device for surface condition evaluation that can be accurately converged.

Yet another object of the present invention is to provide a new and improved apparatus and method for detecting residue on specimens.

Yet another object of the present invention is to provide a new and improved apparatus for evaluating dopant concentration near a specimen surface prior to annealing.

Yet another object of the present invention is to provide a new and improved dope concentration measurement method and apparatus that exceeds the sensitivity of previously available devices.

For these and many other purposes, the present invention provides a new and improved method and apparatus for evaluating the surface condition of a specimen. The method and apparatus are based on the principle that when a specimen is heated, its optical reflectance changes differently depending on the various surface conditions. It is known that optical reflectance depends to some extent on temperature. This dependence is determined by the following equation.

(1) R _T = R _p + (∂R/∂T) (ΔT) In this equation, R _p represents the reflectance at the set temperature, and the second term is the change in reflectance due to temperature change on the sample surface. represents. The term (∂R/∂T) is the temperature coefficient of reflectance and represents the rate of change in reflectance with respect to temperature change. The first term R _p is the temperature change
If it is less than 100 degrees, it is at least 4 digits larger than the second term. Furthermore, the noise level associated with R _p when measured with a photodetector is of the order of √ _p . This value is also about 100 times larger than the second term in the equation, making measurement of the second term very difficult. Expressed in absolute value, the ratio (∂R/∂T) (ΔT)/R _p has a value of 10 ^-4 to 10 ^-5
degree and therefore has never been used as a measurement parameter.

According to the present invention, this difficulty is overcome by modulating the heating source. The periodic changes in reflectance that occur at the frequency of the modulated beam are then monitored. This information is obtained by processing the signal through a narrow band filter. What results is not an absolute reflectance R _T but only a periodic reflectance signal ΔR _T due to the periodic temperature change ΔT.

This periodic reflectance signal ΔR _T is defined by the following equation.

(2) ΔR _T = (∂R/∂T) (ΔT) As can be seen from the above equation, the periodic reflectance signal ΔR _T
is the product of the temperature coefficient of reflectance (∂R/∂T) and the periodic surface temperature ΔT. Given a surface area, the relative effect of these two variables on the change in periodic reflectance (R _T ) depends in part on the thickness of the surface area.
More specifically, where the residual layer (or disturbed surface area) is less than 100 angstroms, the effect on the periodic reflectance signal ΔR _T due to the presence of that surface layer is primarily due to the reflectance This is the result of a change in temperature coefficient (∂R/∂T). This is because such a thin layer has little effect on the periodic surface temperature ΔT.

Changes in the temperature coefficient of reflectance (∂R/∂T) can be measured even for very thin layers. In fact, it has been found that changes in the temperature coefficient of reflectance can be measured even when the layer is as thin as 5 angstroms. This sensitivity is based on the fact that a "clean" or undoped substrate has a certain reflectivity. The presence of a doped material or residual layer having a thickness on the order of 5 angstroms nevertheless measurably changes the temperature coefficient of reflectance. The sharpening ability of such thin layers is based on the fact that reflection of the electromagnetic beam occurs at the specimen surface and that any residue or doping material at or near the surface boundaries will affect the temperature coefficient of reflectance.

Where the residual layer or disturbed doped material region is greater than 100 angstroms, the thermal parameters of the surface layer also affect the periodic reflectance signal _ΔRT . As will be appreciated, this layer has different thermal properties than the layer below it. When this layer is very thin, its thermal properties are almost completely determined by the underlying substrate. However, as the layer thickness increases, its effect on the thermal properties of the system becomes more pronounced through changes to the cyclic surface temperature ΔT.

Where there is a periodic heating source, the varying surface temperature is a result of and is determined by the underlying heat wave. Generation of heat waves by focusing an intensity-modulated heat source onto a specimen is described in earlier U.S. Pat.
No. 4255971. The relationship between reflectance changes and heat wave generation provides an entirely new means for measuring heat waves in specimens. Measurement of heat waves based on reflectance changes of specimens is described in US Pat. No. 4,636,088.
Disclosed in the issue.

Based on the aforementioned principles, very sensitive methods and devices for evaluating the surface condition of specimens can be developed. According to the invention, the apparatus includes a device for generating localized periodic heating on the specimen surface. An optical probe is directed into the region that is periodically heated. An apparatus is provided for detecting changes in the intensity of a probe beam reflected from its surface. As mentioned above, only the changes that result from periodic heating are of significance.

In one preferred embodiment, the specimen is rastered relative to the heating and probing beams. Any change in the output signal of the detector indicates the presence of dopants or residue. Alternatively, the measured intensity changes are compared to known specimens to assess the surface condition.

Objects and advantages of the invention will become apparent from the following detailed description taken in conjunction with the accompanying drawings.

As mentioned above, there is great interest in developing methods and apparatus for evaluating the surface condition of specimens. As shown in FIGS. 1 and 2, there is one pressing problem associated with the semiconductor industry, and that is with the detection of residues produced in lithography and etching processes. The importance of this problem is highlighted by the fact that even small amounts of residue left behind by either the lithographic or etching process can cause circuits to function improperly. The difficulty of solving this problem can be understood by looking at FIGS. 1 and 2. As can be seen, the residue to be detected is in the contacts or holes created in the substrate surface. The width of these holes can be as small as 1 micron. Therefore, no matter what detection method is used, the arrow H in FIGS.
1. It must be possible to look down into these holes as shown in H2. The present invention can overcome these problems by providing a system that can accurately focus into these holes to detect debris.

As mentioned above, another problem relates to the evaluation of the ion dope concentration at the specimen surface, as shown in FIG. The present invention provides a means to solve both of these detection problems.

Referring now to FIG. 4, there is shown an apparatus 20 for carrying out the method of the present invention. In FIG. 4, a specimen 22, which may be an integrated circuit or any other substrate to be evaluated, is shown. The specimen 22 rests on a specimen stage 24 that allows rastering of the specimen to the detection device. Platform control stages are known in the prior art and are also disclosed in the above-mentioned US Pat. No. 4,255,971.

According to the invention, the upper surface of the specimen is subjected to periodic localized heating. In the illustrated embodiment, this periodic localized heating is provided by a laser beam 34 that is intensity modulated by a modulator 32.
The frequency of this modulation is performed by an electronic device, and is preferably 50 KHz or higher.

This intensity modulated heating source may be provided by electromagnetic radiation of various wavelengths including x-rays, gamma rays, infrared, ultraviolet, visible light, microwaves or radio waves. This intensity modulated heat source can also be generated by interaction of the specimen with an intensity modulated stream of particles such as electrons, protons, neutrons ions or molecules. However, it is believed that the illustrated embodiment is preferred because the laser can be focused.

The intensity modulated beam 34 passes through a dichroic mirror 36 and is focused onto the specimen, preferably by a microscope objective 38. In a preferred embodiment, the heating beam is an argon ion laser and the dichroic mirror is transparent to the argon ion radiation.
As discussed below, this dichroic mirror functions to reflect the probe laser beam. Preferably, the probe laser is generated by a helium-neon laser. The periodic heating sources disclosed above are of the type used in various heat wave generation systems. In effect, the intensity modulated beam serves to generate heat waves in the specimen.
As described in the above-mentioned US Pat. No. 4,255,971, information regarding the thermal conductivity of the top layer of the specimen is determined by analysis of the heat waves generated in the specimen. The reflectance signal is influenced by the thermal parameters of the specimen defined by equation (2). Therefore, if the layer of interest is greater than 100 angstroms thick, some component of the reflectance signal ΔR _T will be representative of the thermal properties of the layer. In that case information regarding the layer can be determined by analyzing the heat waves.

Where the layer is relatively thin, its thermal properties do not measurably affect the thermal parameters of the specimen and therefore do not measurably change the cyclic surface temperature ΔT. However, if the layer has a thickness of at least 5 angstroms, this layer changes the temperature coefficient of reflectance (∂R/∂T) of the specimen. This phenomenon is used to the advantage of the present invention, which makes it possible to evaluate very thin layers on the basis of changes in the temperature coefficient of reflectance.

According to the invention, the detection device is defined by an optical probe beam directed onto a sample surface that is periodically heated by a modulated energy beam 34. In the illustrated embodiment, the optical probe beam 5
2 is generated by a helium-neon laser 50. Various other sources of electromagnetic radiation can be used for the probe beam, as long as the reflectance of the beam is measurably influenced by temperature changes at the specimen surface. A probe beam 52 is emitted from a helium-neon laser 50 and then passed through a polarization splitter 5.
Pass 4. In this polarization splitter, coherent light emitted from the laser 50 is transmitted to the splitter 5.
4 is oriented to freely pass through. However, this splitter reflects all light whose phase is rotated by 90 degrees with respect to beam 52. The reason for this configuration will be explained below.

Optical probe 52 then passes through quarter-wave plate 55 . Wave plate 55 functions to rotate the phase of the beam by 45 degrees. As will be appreciated, on the beam's return trip, the wave plate rotates the beam phase by an additional 45 degrees, so that when the beam reaches splitter 54, the beam phase has been phase rotated by a total of 90 degrees compared to when it entered. . This configuration causes splitter 54 to reflect the retroreflected light beam to detector 56. This will be explained in more detail below.

The probe beam 52 first passes through the wave plate and is then reflected downward by the dichroic mirror. As mentioned above, the dichroic mirror is transparent to the argon ion beam, but reflects light in the helium-neon frequency range. In a preferred embodiment, the heating beam and probe beam are aligned such that they are both passed through overlapping lens 38 and focused onto the same small area on the specimen surface. Maximum signal output can be achieved by focusing the probe beam and heating beam onto the same small area.

It should be appreciated that the reflectance signal of interest exists at any region on the specimen surface that is periodically heated by beam 34. Therefore, the probe beam does not need to be correctly overlapped with the heating beam in order to detect the signal of interest. Therefore, microscope objective lens 38 is not required to focus either heating beam 34 or probe beam 52. Rather, it is only necessary to direct the probe beam into at least a portion of the periodically heated region. Discussion and equations for calculating the size of the periodically heated area are found in US Patent Application No. 4,521,188, which is incorporated herein by reference. Generally speaking, the diameter of the heated region extends outward from the center of the heating beam and is a function of the diameter and frequency of the heating beam and the thermal parameters of the specimen.

Since the signal to be measured is very small, on the order of 10 ^-5 of the DC level of the probe beam, every effort must be made to maximize the power for detection. It is therefore desirable to direct the probe beam to overlap the heating beam. Beam movement is achieved by rotating mirror 36. In the preferred embodiment, a small "hole" or "contact" is being inspected, and the optical lens system can focus the incident heating and probe beams into a small area as large as 1 micron in diameter.

The probe beam 52 is incident on the specimen surface and interacts with the electrons thereon and thus with the lattice structure of the specimen. The lattice structure of the specimen changes periodically as the temperature of the specimen changes periodically. The probe beam essentially “sees” this lattice structure
Therefore, the reflection level of the beam changes as the thermal state of the sample surface changes. This probe beam is reflected to a dichroic mirror where it further passes through a quarter wave plate 55 along with the incident beam. As mentioned above, the wave plate 55 rotates the probe beam phase by an additional 45 degrees when the beam reaches the splitter 54, resulting in an overall change to the incident beam.
Rotate the phase by 90°. The splitter therefore reflects the retroreflected probe beam upwardly toward detector 56. Since changes in the intensity of the probe beam are to be detected, a standard photodetector can be used as the sensing mechanism. The measured intensity changes are applied as output signals to a processor 58 to evaluate the surface condition of the specimen.

The operation of processor 58 depends on the type of subject used. In all cases, however, the processor is designed to evaluate periodic intensity changes of the incident probe, which are manifestations of reflectance changes caused by periodic heating on the specimen. These high frequency changes are filtered to produce a signal that can be evaluated.

For simple small area inspections, the processor is programmed with known signal values, such as uncoated or raw oxide. If a residue is present, for example as shown in hole H1 in FIG. 1, the measured output signal will differ from the known stored signal value, indicating the presence of the residue. The evaluation of the thickness of the residue layer is very difficult since the specific residue type is not known in most cases.
However, most manufacturing processes only require detection of the presence of residue.

Another use of this system is to evaluate complete integrated circuits. In this case, the specimen can be rastered to the heating and probe beams. For example, if the movement of the specimen stage 24
The program can be programmed to follow the paths defined by the etched tracks shown in the figure. If there is no residue, the output signal will remain constant. On the contrary, any peak value or change in the output signal indicates the presence of a residue.

More quantitative information is available for ion-implanted dopes. The reason is that because of the identity of both the substrate and the dopant, the thermal properties are known. For example, the output signal can be normalized to a known reference sample to obtain the actual implanted dope concentration. As mentioned above, this system can detect doping levels prior to the annealing process. Unlike electrical testing, which requires the dope to be activated, ions that are fixed in interstitial positions before annealing affect the reflectance temperature coefficient and can therefore be monitored.

Furthermore, in practice, the present invention has been shown to be capable of detecting boron ion dopant concentrations in silicon down to as low as 10 ¹⁰ ions/cm ² . This sensitivity is three orders of magnitude greater than electrical tests available in the prior art.
By rastering the sample relative to the heating and probing beams, analysis is also possible to create a one-dimensional or two-dimensional map of dope concentration.

In summary, a new and improved method and apparatus for evaluating the surface condition of a specimen is provided. The present invention provides an apparatus for providing localized periodic heating of a specimen surface. In a preferred embodiment, the heating device is an intensity modulated laser. The detection system is further focused onto the region which is periodically heated by the intensity modulated laser as reflected from the specimen surface. Apparatus is provided for detecting changes in the intensity of this reflected light as a result of this periodic heating. By processing the measured intensity changes of this reflected beam, the surface condition of the specimen can be evaluated.

Although the invention has been described in terms of preferred embodiments, those skilled in the art may make various other changes and modifications without departing from the scope and spirit of the invention as defined by the claims.

[Brief explanation of drawings]

1 and 2 are cross-sectional views illustrating typical integrated circuit lithography and etching processes and problems associated therewith, and FIG. 3 is a cross-sectional view of an integrated circuit showing dope ion implantation; Fourth
The figure is a combination of block diagram and schematic diagram of an apparatus for carrying out residue detection according to the invention. 20... Device of the present invention, 22... Sample, 24... Sample stage, 30... Heating beam generator, 50... Probe beam generator, 34... Heating beam, 52... Probe beam, 32... Heating beam modulator, 36...
Dichroic mirror, 38... Converging lens, 54... Polarization splitter, 55... 1/4 wavelength plate, 56... Probe beam intensity change detection device, 58... Processing device for detected intensity change.

Claims (1)

[Claims] 1. An apparatus for evaluating the surface condition of a specimen, comprising: an apparatus for inducing local periodic heating on the specimen surface; a radiation probe beam; and a radiation probe beam reflected from the specimen surface. a device for directing the radiation probe beam onto an area periodically heated by the periodic heating device, such that a device for measuring periodic intensity changes of the reflected radiation probe beam; and processing the measured periodic intensity changes of the reflected radiation probe beam generated at a modulation frequency of the periodic heating device to determine the periodic intensity changes of the specimen. A device for evaluating a surface condition; and an evaluation device comprising: 2. The evaluation device according to claim 1, wherein the device for directing the radiation probe beam is configured to direct the probe beam to the center of a region periodically heated by the periodic heating device. evaluation equipment. 3. Evaluation device according to claim 1, in which the periodic heating device is defined by an intensity-modulated laser beam. 4. The evaluation device according to claim 3, wherein the radiation probe beam is directed so as to overlap with the laser beam. 5. The evaluation device according to claim 4, further comprising a device for converging the heating laser beam and the radiation probe beam into a small area with a diameter of 1 micron. 6. The evaluation device according to claim 1, wherein the device for measuring periodic intensity changes of the reflected radiation probe beam is a photodetector. 7. An evaluation device according to claim 1, wherein the processing device functions by comparing periodic intensity changes of the reflected radiation probe beam with known reference intensity changes. 8. The evaluation device according to claim 1, further comprising a device for rastering the specimen with respect to the heating device and the radiation probe beam, and the processing device performs rastering on the specimen surface. An evaluation device operative to compare periodic intensity variations of said reflected radiation probe beam. 9. The evaluation device according to claim 1, wherein the radiation probe beam is determined by a laser. 10 A method for evaluating the surface condition of a specimen, comprising:
heating a specimen surface periodically; directing a radiation probe beam to be reflected from the specimen surface onto a portion of the region that is periodically heated by the periodic heating; measuring periodic intensity changes of the reflected radiation probe beam caused by heating-induced periodic changes in the optical reflectance of the specimen; and the measurements occurring at the modulation frequency of the periodic heating. and processing periodic intensity changes of the reflected radiation probe beam to evaluate the surface state of the specimen surface. 11. The method of claim 10, wherein the radiation probe beam is directed to the center of a region that is being periodically heated by the periodic heating. 12. The method of claim 10, wherein the step of evaluating is performed by comparing periodic intensity variations of the reflected radiation probe beam with known intensities. 13. The method of claim 10, further comprising the step of performing specimen rastering relative to the heating, and wherein the processing step comprises the step of: A method comprising comparing periodic intensity changes of a beam. 14 A method for evaluating the surface condition of a specimen, comprising:
directing an intensity-modulated laser beam to periodically excite localized regions on the surface of a specimen; generating a radiation probe beam; and focusing the laser beam and the probe beam to overlapping substantially on the surface of a specimen; and measuring periodic intensity changes of the radiation probe beam based on changes in optical reflectance of the specimen after reflection from the surface of the specimen; processing the measured periodic intensity changes occurring at the modulation frequency of a modulated laser beam to evaluate the surface condition of the specimen.